Platform for Martin Michlmayr

Background

My name is Martin Michlmayr. I hope that you will elect me as the next Debian Project Leader because I am the person best suited to motivating and coordinating the Debian community.

I hold a Master degree in Philosophy and have recently completed a Master of Science in Psychology. I'm currently doing a Master of Software Systems Engineering at the University of Melbourne and am looking forward to pursuing a PhD in Software Engineering about Debian and Free Software afterwards. I have already carried out some research in this regard and I'm currently working on a paper about Debian together with the anthropologist Biella Coleman and with my colleague Mako Hill.

I have participated in various Free Software projects for over 8 years. In 1994, I joined the GNUstep project which is working on a free implementation of NeXT's (now Apple's) object-oriented development environment OpenStep. My main task involved the coordination of volunteers who were interested in GNUstep. Later, I joined Linux International as its Publicity Director and Membership Secretary. In this function, I worked together with Jon "maddog" Hall and other important figures of the Linux Community.

In addition to my involvement in various Free Software projects, I have also held several jobs related to Free Software. I acted as the editor of a German Linux magazine twice and managed to recruit famous developers such as Stephen van den Berg (author of procmail) and Matthias Ettrich (founder of KDE). Additionally, I have published several articles about Free Software projects. Those who are interested in the full story may check out my resume.

My Involvement in Debian

I joined Debian in October 2000. Considering the age of the project I may be perceived as a fairly "new" developer; yet, due to my work, I'm considered to be a "senior" developer by many.

After joining Debian, I decided to help out with the New Maintainer (NM) process. I became an Application Manager (AM) and helped many people interested in Debian join the project. I have been a very successful AM, completing more than 100 and rejecting more than 50 applications. Later, I joined the Front Desk and since then have helped to coordinate the NM process. I assign applicants to AMs, make sure that their application process is going smoothly and act as a bridge between applicants, AMs and the Debian Account Manager (DAM).

In addition to helping people to join Debian, I'm also trying to find inactive developers. They are becoming a problem since they are often responsible for a package but don't maintain it properly. I contact developers who might be inactive periodically and store information gathered during this process in a dedicated database so other developers can follow and help with this work. When I have concluded that a developer is in fact inactive, I usually take away their packages so active developers can take over and actually maintain the packages.

I'm also involved in Debian's Quality Assurance (QA) effort. For example, I have created web sites that show the status of the base and standard packages at a glance. These pages have been very helpful during Bug Squashing Parties. Those pages have now been integrated into our Bug Tracking System (BTS) properly by Colin Watson.

Finally, I created the GPG Key Signing Coordination page so prospective developers can easily find a Debian Developer in their area.

Most importantly, I work hard to let work on separate aspects of the project inform and integrate. I am able to be successful because each of these projects is an essential and interconnected piece of a larger whole and because I'm able to approach small tasks from this perspective. Through the NM process, I gather information that is very important while searching for inactive maintainers. Also, finding an inactive developer is important for QA. Similarly, a buggy package found during QA work prompts me to check whether the maintainer is still active. It is interesting to note that the connections between these tasks can often be seen in my work. For example, I found at least one developer for whom I have acted as Application Manager and who is inactive now. Also, there are several developers who were my applicants who have in the meantime resigned from Debian. It is always strange to see people come and go, but I suspect it's a natural process in Free Software projects.

Furthermore, I try to integrate information from other projects, such as GNU to which I have fairly good connections. Quite recently, I have worked with GNU to locate an inactive maintainer who is working on Debian as well as GNU.

As you can see, it is very important for my work to make connections between different parts of the projects. While different tasks may appear to be distinct on the surface, they are usually strongly connected with each other. This is also the reason why I decided to study Philosophy and Psychology instead of Computer Science. I want to see how the different areas interact, and I usually profit in one area from the others. Seeing these connections usually allows me to coordinate tasks very well. I have done coordination tasks for various projects and I think the way my memory works is part of the reason why I'm fairly good with this -- I can keep many different pieces of information in my memory at the same time.

The Role of the DPL

Unfortunately, the role that the DPL should play is not well defined. Debian is a very large projects and different people have different expectations of the DPL. I have found Branden's questionnaire very stimulating and I hope there will be more discussions about the expectations developers and users have from the DPL in the future.

Similarly, while everyone complains about the release cycles of Debian, it is not clear at all how frequently we should release. For example, SuSE and Red Hat have introduced offerings of their distributions that have similar release cycles to that of Debian. There are many reasons why you can upgrade your system only every two years. Yet, there are also users who want the newest software and like to upgrade twice a year. Finding a balance between the different requirements will not be easy. We really have to actually identify the requirements our users have. The poll by Joey about security updates for potato has been a first step which I applaud.

External Functions

One of the DPL's responsibilities is to represent Debian to the outside world. As such, he should give talks and work on collaborating with other projects.

I am a fairly experienced speaker, having given several talks about Debian. For example, I have given presentations about Debian QA at the Debian Conference in Bordeaux, LinuxTag in Stuttgart and at the Free Software Symposium in Tokyo. I have also given a talk about our archive maintenance tools at LinuxTag in Karlsruhe and have recently talked about adapting Debian to embedded use at FOSDEM in Brussels.

I intend to give more talks in the future. In particular, I plan to attend LinuxTag in Karlsruhe and DebCamp and DebConf in Oslo. I'm also interested in speaking at other conferences. However, I can certainly not attend all conferences because, as a student, I have classes to attend and have a limited budget. In the latter case, I have done fairly well up to now since I travel quite cheaply. I tend to stay at other people's places and have therefore low expenses. I have traveled through five countries in the last week and this would certainly not have been possible without the community. Also, I would like to thank the organizers of the Free Software Symposium in Tokyo for inviting me and paying my expenses as well as Linux International for a grant to attend the Debian Conference in Bordeaux.

Still, I imagine that I will receive more offers for talks than I can fulfill. It would be a shame to waste those opportunities. Thus, I intend to coordinate with other developers who are capable of giving talks. For example, if Bdale's employer is willing to continue sponsoring his travel to give talks, I might make him an official representative of Debian and coordinate with him. Furthermore, I will work with the Debian Events people to get a listing of experienced speakers on our web site. This will help organizers of conferences find people who can give talks about Debian.

Also, I will try to set up more partnerships. For example, I think that Debian should work together more closely with hardware vendors and get a certification program set up. Many argue that a volunteer project cannot do that, but I see no reason why it should not be possible. Furthermore, Debian could apply for funding as part of EU's 6th Framework Programme. In fact, I spoke to Peter Vandenabeele of Mind.be and he would be willing to coordinate this.

We should also try to cooperate more with projects who use Debian and enhance it. These efforts and results should be integrated into Debian. A good example where this works is Melbourne University's Trinity College. They are working on packages for the new XFree86 release and are working together with the XFree86 maintainer. However, there are many efforts where the results are never given back to Debian, which is a waste of effort. I'd also love to see those users who need security updates become more active and get involved by helping out.

Finally, as DPL I will represent the Debian Project to Software in the Public Interest (SPI). I have been an advisor of SPI for over 1.5 years and I'm part of its Membership Committee. Hence, I know how SPI works. There is one major endeavor I intend to pursue together with SPI and its treasurer. There are many individual Debian Developers in different countries who posses money that is dedicated for Debian. This money comes from selling t-shirts at conferences and similar activities. I would like to see a listing created of who keeps how much money for Debian. This list will be very useful. For example, when a Debian Developer in a specific country needs money to organize an event we can check whether there is any money in that country. I think it does not make sense to have local SPI branches in every country. Instead, we should try to work together with the FSF who is creating non-profit organizations in many countries.

Internal Functions

While the external functions are very important, I think the internal functions should be the DPL's first priority. There are several endeavors I would like to pursue. First, I think there are several problems with the sponsorship system. There is no listing of sponsored people and hence it is quite hard to keep track of them. A recent posting to the debian-qa group showed that there are about 200 sponsored people. I wonder how many of those 200 people are not active anymore. Unfortunately, we don't have the same tracking system we have for Debian Developers (echelon) for sponsored people. However, a listing of sponsored people would allow us to change this. I will therefore propose to have sponsorship integrated with the NM database as a pre-stage of NM. People will be able to sign up for sponsorship and then transfer to NM. Another feature of this is that we have a listing of people who are still looking for a sponsor.

Second, I will re-introduce the New Maintainer postings. There are so many Debian Developers that you don't know all of them. In the past, a short description of a developer was posted when they joined the project. I will post such summaries to debian-project and archive them on a web site. Of course, in order to have such postings, we actually need to have applicants approved by the Debian Account Manager (DAM). This has been problematic in recent months. However, since I have worked on New Maintainer for such a long time, I'm certainly in a good position to help getting this fixed. In fact, I spent last weekend in Cambridge with James Troup going through pending applications. You should see results from this soon.

Third, I intend to put more focus on inactive developers. They are becoming an increasing problem and the DAM has also shown interest in doing something about them. In contrast to previous DPL candidates who were running for DPL partly to gain the authority to do something about inactive maintainers, I have actually done significant work in this area already. I have tracked down many inactive developers and have orphaned almost 300 packages. Due to this work, I am now perceived as the authority in this area.

The last point is very important. No one has given me the power to orphan packages. I have simply done it. And since I was very careful about it, I have received almost no complaints and most people see my work as very valuable. As DPL, I will encourage similar efforts. You don't have to be DPL to get things done. Anyone can do Good Stuff!

So why am I running for Project Leader anyway? All of the three intended projects can be done without being DPL (and as an added bonus, I will do them regardless of whether I am elected or not). This leads me back to the question of the role and power of the DPL.

I think one of the main tasks of the Project Leader is to coordinate and motivate people - to lead. This is why I said that while the external functions of the DPL are important, the internal functions are even more important. While it is quite hard to lead a project consisting of so many people with so diverse expectations and personalities, I think that it can actually be done. Thus, my main aim as Debian Project Leader is to lead, motivate and coordinate.

The most important thing of Debian is its community. We would be nowhere without our great community and with the Free Software community at large. I will therefore pursue actions to increase the feeling of a community. For example, I think that meetings in real life such as Debian Conference are very important, and I will help getting sponsors so many Debian Developers can attend. Future interactions are just different once you have gotten drunk with someone... Furthermore, I will try to create the infrastructure so people can put pictures and short biographies of themselves online (only if they want to). This should help give a less impersonal feel to the names we see on mailing lists. It is a shame that three different DPLs have done nothing about #76187 and that I actually have to run for DPL myself to get this fixed...

Another factor to consider is that we are part of a larger community. I think it's very important to have good contacts to upstream maintainers. As such, you should try to encourage your upstream to attend DebCamp, and I hope we can get some funding for this.

It is often claimed that Free Software projects are not managed. And while it is true that they are not managed in a traditional way, there is a great deal of coordination going on. There are companies who pay much money to understand how the Free Software community works. I intend to lead and coordinate. However, I will not do any Bruce-Perens-style management since this is not the way Debian works these days. I am aware that Debian is a very diverse project and that every developer has to be treated individually. For some people, it's best if they are just left alone to do their stuff. Others, on the other side, might need someone to tell them what to do. It's crucial for the Project Leader to find out who has to be treated in which way and act accordingly. As a result of the traveling I have done recently, I have met many other developers. Knowing people in real life certainly helps when it comes to leading the project.

Also, I encourage other people to help with coordination tasks. While it's good to have one Project Leader and one Release Manager (RM), everyone who prompts someone to get work done is acting as a mini-DPL or mini-RM and we should encourage it. When you see that there is an RC bug and you know someone who might be able to fix it, then ask them nicely. It's no use dictating anything in Debian since we are a volunteer project. However, asking people nicely in private mail usually helps a lot. Hence, if you see an opportunity to coordinate, then go ahead and do it.

Conclusions

While I intend to represent Debian to the outside world, my main focus will be on Debian internally. We need a leader. Someone who motivates people and who coordinates. Furthermore, the Project Leader has to be visible inside the project.

We are part of a great movement. We are actually changing the world with the Free Software movement and with Debian. I will do anything that helps to make this movement an even greater success. I think I can best employ my abilities by coordinating, motivating and leading the members of Debian.

Rebuttal

Moshe Zadka

Moshe's intention is to do nothing at all since the project manages itself. Unfortunately, with this attitude, we would not go anywhere. If everyone thought they would not have to do a specific task because someone else might do it, then things will never get done. There is a lot of coordination going on in the project. This coordination is crucial for the project, and it's the job of the Project Leader to coordinate these coordination efforts.

Bdale Garbee

I found it interesting that Bdale speaks of communication in his platform because lack of communication and visibility in the project is the reason of my disappointment. While I have heard that Bdale has done a huge amount of communication behind the scenes which was very important in getting things (such as keyring) fixed, I personally felt that the community at large was not well informed at all of what was going on.

Branden Robinson

I didn't see many new thoughts in Branden's platform that were not discussed in previous years. In general, I doubt that Branden would actually be able to lead the project. This is partly due to the fact that he has had (pointless) arguments with virtually anyone in the project who is doing important stuff (with Anthony Towns on the RCness of a bug, with the FTP masters on replacing .orig.tar files, etc). While these arguments of the past do not mean that he cannot work together with those people now, I think that it might make it more difficult. Also, this should not be an ad hominem attack, but one may not forget that the DPL is a social position. I respect Branden for his work on the XFree86 packages and for PGI, but I doubt he would be an effective leader.